Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
73 www.erpublication.org
Rainfall Runoff Simulation of Shipra river basin using AWBM RRL Toolkit
74 www.erpublication.org
International Journal of Engineering and Technical Research (IJETR)
ISSN: 2321-0869 (O) 2454-4698 (P), Volume-5, Issue-3, July 2016
E. ACCURACY CRITERIA Table 1: Optimized parameters of RRL AWBM model
Parameter Optimized Range of the
Accuracy of the model can be examined on the basis of
Parameter Values Parameters
Coefficient of Determination (R2), Efficiency Index (EI) and
Root Mean Square of Error (RMSE). The use of the A1 0.134 0.000 1.000
coefficient of determination (R2) is to test the goodness of fit
of the model and explains predictive power of model. A2 0.433 0.000 1.000
BFI 0.494 0.000 1.000
C1 29.85 0 50
Where, qo= observed flow, qo= mean value of observed flow, C2 116.90 0 200
qs= simulated flow and n = number of data points. C3 193.67 0 500
The reliability of the model was evaluated on the basis of Kbase 1.00 0.000 1.000
Efficiency Index (EI) as described by the Nash and Sutcliffe
[13]
Ksurf 0.252 0.000 1.000
The RRL AWBM model was setup to carry out Fig. 4: Comparison between observed discharge and
Rainfall-runoff modeling in Shipra river basin at Ujjain G/d simulated discharge during the calibration of AWBM model
site having catchment area 2102 km2. The AWBM model was for specific year 1998.
calibrated for six years period from 1996 to 2001 and then
validated for the remaining period of five years from 2002 to
2006. Optimized parameters of RRL AWBM model found
using genetic algorithm. The graph presenting comparison
between observed and simulated discharge during model
calibration is shown in Fig. 3 which gives the idea and view of
best match obtained during the model calibration. Fig. 4
shows the comparison between observed and simulated
discharge during the calibration of AWBM model for specific
year 1998 indicates the very good match. Similarly Fig. 5 and
Fig. 6 show the comparison between observed discharge and
simulated discharge during the validation of AWBM model.
The observed and simulated hydrographs were found Fig. 5: Comparison between observed and simulated
matching well for peak and low flows as well reasonably. discharge for validation
75 www.erpublication.org
Rainfall Runoff Simulation of Shipra river basin using AWBM RRL Toolkit
V. CONCLUSION
In this study, AWBM RRL model has been developed and
tested for the performance and suitability in Shipra river basin
of Madhya Pradesh, India. AWBM is lumped and conceptual
rainfall-runoff model developed by CRCCH, Australia and
available in public domain. AWBM model was calibrated and
validated using daily weighted precipitation, daily
evapotranspiration and daily observed runoff time series of 11
years period from 1996 to 2006. The model was calibrated for
the six years period from 1996 to 2001 and validated for five
Fig. 6: Comparison between observed discharge and years period from 2002 to 2006. Genetic algorithm was used
simulated discharge during the validation of AWBM model for the optimization of the model parameters of the AWBM
for specific year 2003 model. The sensitivity analysis was carried out on all the
AWBM parameters so as to find the sensitive parameters. The
The Performance of AWBM model was evaluated based on AWBM model was tested evaluated based on Accuracy
Accuracy criteria such as Coefficient of determination (R2), criteria such as Coefficient of determination (R2), Efficiency
NashSutcliffe Efficiency Index (EI), Root Mean Square Index (EI), Root Mean Square Error (RMSE), and correlation
Error (RMSE), and correlation coefficient (r). The coefficient coefficient (r). The coefficient of determination R2 value for
of determination R2 value for calibration and validation calibration and validation period is 0.656 and 0.496
period is 0.656 and 0.496 respectively which indicating good respectively which indicating good agreement between the
agreement between the observed and simulated runoff. The observed and simulated runoff. The NashSutcliffe
NashSutcliffe Efficiency (E) for calibration and validation is Efficiency (E) for calibration and validation is 65.40% and
65.40% and 48.40% respectively. The value of these accuracy 48.40% respectively. The model was found suitable for
parameters indicates that the AWBM model is performing Shipra basin in simulating hydrological response of the basin
well in predicting runoff. to the rainfall and predicting daily runoff with good degree of
Table 3: Accuracy Parameter values during calibration, accuracy.
validation and total period REFERENCES
Total
Accuracy criteria Calibration Validation [1] Linsley, (1982) Rainfall-runoff models - An overview in
period rainfall-runoff relationship, Proceedings of the International
Symposium on Rainfall-Runoff Modeling, May 18-21, 3-22.
Coefficient of
0.656 0.496 0.622 [2] Lipiwattanakarn, S, Sriwongsitanon N, Saengsawang S.(2004)
determination (R2) Improving neural network model performance in runoff estimation by
using an antecedent precipitation index. Journal of Hydroscience,
Nash Sutcliff 22(2), 141-154.
65.40 48.40 60.40
efficiency (EI) % [3] Shoemaker, L., Lahlou, M., Bryer, M. D. & Kratt, K. (1997)
Compendium of tools for watershed assessment and TMDL
Correlation development. U.S. Environmental Protection Agency (EPA),
0.810 0.790 0.789
coefficient (r) 841-B-97-006.
[4] Galkate, R. V, Jaiswal, R. K., Thomas, T., Nayak T.R., (2011)
Root mean square Rainfall Runoff Modeling Using Conceptual NAM Model India
58.40 56.74 57.65 Madhya Pradesh Map of Bina Basin up to Rahatgarh Site, Institute of
error (RMSE)
Management and Technology, Nagpur.
The sensitivity analysis of AWBM model parameters was [5] CRC for Catchment Hydrology (2004), Rainfall Runoff Library User
performed so as to understand how sensitive a model is to Guide, 1.0.5, Australia.
[6] Sharifi, F and Boyd, M.J. (1994), A Comparison of the SFB and
certain parameters. The AWBM provides a feature to AWBM Rainfall-Runoff Models, Surface Hydrology and Water
examine the sensitivity of model parameters. Sensitivity Resources Papers, 15, 491-494.
analysis indicates that four parameters are sensitive and four [7] Boughton, W. (2006), Calibration of Daily Rainfall Runoff Model
are non sensitive. with Poor Quality Data, Environmental Modeling and Software,
Elsevier, 21(8), 11141128.
Table 2: Sensitivity analysis of AWBM model parameters [8] kumar, P. and Hanuma Rishi, K. (2013), Simulation of Rainfall
Runoff using SCS & RRL (Case Study Tadepalli Mandal),
Model Parameters Sensitivity Analysis International Journal of Engineering Research and General
Science,1(1), 23-33.
A1 Non sensitive [9] Balvanshi, A. and Tiwari, H.L. (2015), Rainfall Runoff Estimation
Using RRL Toolkit, International Journal of Engineering Research &
A2 Non sensitive Technology, 4(5), 595-599.
[10] Jamal, J.F.,(2011), Comparison of Conceptual and Neural Network
BFI Sensitive Models for Daily Rainfall-Runoff Modelling, ICCEES, Pattaya.
[11] Haque, MD., Rahman A., Hagare, D. and Kibria, G. (2015),
C1 Non sensitive Parameter Uncertainty of the AWBM Model when Applied to an
Ungauged Catchment, Hydrological Processes, 29(6), 14931504.
C2 Non sensitive [12] Yu, B. (2015), How would peak rainfall intensity affect runoff
predictions using conceptual water balance models? International
C3 Sensitive Association of Hydrological Sciences, 371, 109-115.
[13] Nash, J.E.,(1958) Determination of runoff from rainfall, Institute of
Kbase Sensitive Civil Engineering, Proc., vol. 10, pp. 163-184, 1958.
Ksurf Sensitive
76 www.erpublication.org